Cloud Apps &
Microservices

Microservices or microservices architecture is a design approach aiming to build scalable and highly maintained applications. That is, an application is built with a collection of microservices that are loosely coupled. That’s why, microservices architecture transforms your business agility with independent deployable and testable components.

Traditionally, a large team of developers are working towards shipping a single application. This involves collaborating in a common cycle of build, test and release. With the large number of codependents, deploying new changes can take up to several weeks and months. To put it in another way, one feature change is impacting the build of the whole application.

On the other hand, an application is broken into smaller services in a microservices architecture. These services have a bounded context and they are loosely coupled. Hence, autonomous from one another. Each microservice is managed by a small team.

This team has an end-to-end control over this service in terms of UI, business logic and database. Team members can build, test and release much faster builds without dependencies. If there is an issue with one feature in the application, the issue is not propagating across other modules. Consequently, it doesn’t cause downtime in the system.

Challenges

A leading Mortgage Financing institution, serving citizens on low and middle income, was having several challenges with its existing legacy core system. The legacy system is more than 20 years old. Eventually, it kept growing over the years by adding new features incrementally serving the continuous business demand of the organization.

The system grew tremendously till it reached a stage where adding new features and changing existing ones was very costly, risky & time-consuming. Here is a summary of their challenges:

• Not responsive to business demand: The current system was too complicated to add new features to it. Implementing new features was time-consuming in terms of development, testing and release time.
• Change was risky: The current legacy code was complex to implement new changes as well as testing was time-consuming and risky.
• Scalability: The current system was processing hundreds of thousands of transactions and database contained millions of records. It was difficult scaling the system during peak times and sudden load.
• APIs & Integration: exposing APIs and integrating with external systems was challenging, complex and time-consuming.

Solution & Approach

After assessing the situation we decided to migrate the system to microservices architecture running on containers on a private cloud container orchestration environment.

Solution highlights

• Phased Agile Approach: we started breaking down the legacy system into smaller components “microservices”, migrating these components to containers and exposing them as REST APIs. This was a low risk approach where the 2 systems coexisted together migrating components every 2 – 4 weeks sprints.
• Quick time to market: Implementing new features and deploying it to production improved by 4-5x faster. All new components had automated DevOps pipelines with built test automation; thus achieving quality as speed.
• Performance Boost: We moved components with scalability issues and performance bottlenecks; hence offloading the legacy system to the modern scalable platform. Each microservice can now scale independently based on its load at runtime, and we don’t have to scale the whole system.
• API first: Easily integrate with internal systems like digital channels and external 3rd party systems and banks.